What Affects the Quality of Score Transformations? Potential Issues in True-Score Equating Using the Partial Credit...

Area of Science:

• Psychometrics
• Educational Measurement
• Health Research Methodology

Background:

• Scale equating is crucial for comparing scores across different assessments.
• Concurrent calibration with the partial credit model is a common equating method.
• Understanding factors affecting score transformation precision is vital for reliable measurement.

Purpose of the Study:

• To investigate how construct similarity and scale targeting impact score transformation precision.
• To evaluate the effect of item difficulty differences on scale equating.
• To provide practical insights for scale equating in health-related research.

Main Methods:

• Simulated data for two scales using a common person design.
• Varied construct similarity (factor correlation), item difficulty, and parameter dispersion.
• Employed concurrent calibration with the partial credit model.
• Compared true-score equating with identity equating.

Main Results:

• Low construct similarity significantly reduced score transformation precision.
• Good scale targeting enhanced precision, particularly with decreasing similarity.
• Difficulty shifts up to two logits showed minimal impact on transformation precision.
• Larger sample sizes and longer scales yielded marginal improvements in precision.

Conclusions:

• Construct similarity and scale targeting are key determinants of score transformation precision.
• True-score equating demonstrates robustness against moderate difficulty shifts.
• Findings offer guidance for optimizing scale equating procedures in health research.